# Actively learning to learn causal relationships

arXiv: https://arxiv.org/abs/2206.09777 · HTML: https://mlsanity.com/paper/2206.09777/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 76 references that are in the map.
- First posted on arXiv 20 Jun 2022.

## Builds on (most cited first, 6 of 6)

- [Meta-Learning in Neural Networks: A Survey](https://mlsanity.com/paper/2004.05439/) (2020)
- [Learning Invariant Representations for Reinforcement Learning without Reconstruction](https://mlsanity.com/paper/2006.10742/) (2020)
- [Implicit Deep Adaptive Design: Policy-Based Experimental Design without Likelihoods](https://mlsanity.com/paper/2111.02329/) (2021)
- [Alchemy: A benchmark and analysis toolkit for meta-reinforcement learning agents](https://mlsanity.com/paper/2102.02926/) (2021)
- [Formalizing Neurath's Ship: Approximate Algorithms for Online Causal Learning](https://mlsanity.com/paper/1609.04212/) (2016)
- [Learning Causal Overhypotheses through Exploration in Children and Computational Models](https://mlsanity.com/paper/2202.10430/) (2022)

## Built on by (most cited first, 2 of 2)

- [Disentangled Representations for Causal Cognition](https://mlsanity.com/paper/2407.00744/) (2024)
- [Active Learning Methods for Efficient Data Utilization and Model Performance Enhancement](https://mlsanity.com/paper/2504.16136/) (2025)

## Abstract (arXiv)

How do people actively learn to learn? That is, how and when do people choose actions that facilitate long-term learning and choosing future actions that are more informative? We explore these questions in the domain of active causal learning. We propose a hierarchical Bayesian model that goes beyond past models by predicting that people pursue information not only about the causal relationship at hand but also about causal overhypothesesunicodex2014abstract beliefs about causal relationships that span multiple situations and constrain how we learn the specifics in each situation. In two active "blicket detector" experiments with 14 between-subjects manipulations, our model was supported by both qualitative trends in participant behavior and an individual-differences-based model comparison. Our results suggest when there are abstract similarities across active causal learning problems, people readily learn and transfer overhypotheses about these similarities. Moreover, people exploit these overhypotheses to facilitate long-term active learning.